Planet

Planet @ Danish 2006

-------------------------------

This summary supposes that one has read:

the DFOSC Manual and the BIAS manual at the danish telescope

The different pages of this website

The README files on "ccsh" and the way to make flats in the CCSH directory

The TCS is the same, the acquisition system is BIAS on the acquisition PC.

The images read by BIAS transit through ins1d5 (midas screen), which send them to the "pipeline PC" w1d5dlr.

TCS Computer:

> no more "empty fields" catalog available

> planet objects in /home/dfoscobs/catal/planet06.cat

> Also in /catal, we use to keep in the file planet_guide a list of the x and y values for the guide probes, to avoid searching for a guiding star on every pointing. Keeping a list of this file in a window, and cuting/pasting the values to the guide probe x,y fields at preset time saves time.

Warning: The TCS program disables catalog selection when the /catal directory is modified in any way while it is running !

If you want to update any star/guideprobe catalog, do it before launching the TCS program, initializing, etc...

BIAS Computer.

NAME: pc2-1d5.ls.eso.org 134.171.143.21

> Bias is "line oriented". There is a manuel in the control room. We use to run pre-edited files (batch files), for flats, test, and science. Each file defines the filter, the object name, and the exposure time. We use to keep all the files opened in workspace 2 with emacs. Each file is 3 lines long and defines OBJECT name, filter, exposure time.

A typical file, OB04250R.bach, would read:

OBJECT ZOB06207I

bfilter 5

exp 300

( Do not forget a R/C at end of file, otherwise the last line will not be executed)

In practise, in 2006 so far we didn't use the .bach files and instead type the commands

at the prompt.

All files in: /home/dfosc/data with ext. .bach

note: The BIAS system takes its input in the data directory, where all images are stored.

From time to time, this dir. must be cleaned for dfsc.mt files, and some bad things may happen to the bach files...

It is our responsibility to keep backup up to date !

Some bach files :

start.bat run at cold start and defines the sub-image 512x512 pixel we use, the amplifier, etc

Amp is A, with high gain (standard). Our stamp is:

xbegin 820

xsize 512

ybegin 800

ysize 512

THIS IS THE OFFICIAL PLANET STAMP ON THE CCD. It has been chosen to have minimum defects. The cross on the TV for CorCo (labelled 1050 1050, left of the box in the middle of the screen) corresponds to the center of the planet stamp

Default values (full ccd ) are xbeg 1 xsize 2148 ybeg 1 ysize 2102

For a new file, you may open a new window from a termenal window in /home/dfosc/data using:

/usr/bin/emacs -bg black -fg white -g 35x7 XXXX.bach &

then filling it and saving it before use.

. . . and all the individual run files for BIAS, one per object/filter.

MIDAS COMPUTER (INS1D5) -

It receives images from BIAS, display it with Midas, and send it to the old PC,

w1d5drl, and the pipeline PC w1d5pl

it must be working to transmit the BIAS files to the PCs: w1d5drl and w1d5pl.

seeing midas does not work.

You may set in options: image size to 2,2 to get a decent image...

Focus pyramid is not available. But you can trust the "focus formula" for B,V,R:

focus = 15220 + 12 ( Tserrurier -10 )

Tserrurier is displayed in a little window in the upper right corner of the rightest screen of ins1d5.

Setting the focus is done via the BIAS command: telfocus <value>

or using the yellow box wich displays it on the rack..

IMPORTANT: From time to time, when taking a series of short images (bias & flats ), some of them are lost. John Pritchard is working on it, but you have no choice but to transfer

them by hand

W1d5pl - PC PIPELINE

guest is the working account. we use the disk /scratch which has more space than

the system disk.

It is the only computer which can run mozilla.

Generaly, WorkSpace 1 is kept for Import/pipeline windows.

Rebooting is done via the "red hat" menu.

IMAGE STORY

> An image acquired on BIAS is stored in /home/dfosc/data/dfscxxx.mt

If you need to delete some ( a message displays warning ), make sure to delete the oldest ones.

> Then sent to ins1d5 for display by midas, which

sends it to w1d5drl (only for backup reasons) and to w1d5pl (for processing).

> w1d5drl , in the directory

/scratch/guest/DFOSC_FASU/<date> <date> in the form: 2004-06-09

This directory is the Import directory for planet.

> w1d5pl , in the directory $RAW_IMAGE,

correcting at the same time the header for MJD-OBS, PL_DATE.

The image in the import directory is flagged as ".done".

csh_loop_flatfield bias21jun Domeflat21jun

Flat fielding of science images.

The image is transferred to $IN_IMAGE with the name from the "OBJECT" keyword given in BIAS.

The image in $RAW_IMAGE is set to ".done" if science.

About making bias and flats: refer to $SCRIPT_CODE/README_FLATS.txt

Processing: Is done with ccsh functions.

See the README_CCSH for details.

The steps are: fwhm (sky parameters), Thresh (shallow run), cross (mapping

transformation between the image and the reference "template" frame), deep Dophot

run, and Archiving.

One may run he pipeline step by step, or use "looping" functions (see below).

Sending archives to Planet:

SyncArchives:

a) Synchronize archives with planet@chianti.iap.fr

b) runs "remlog", sending the list of taken images.

24 HOURS FROM THE LIFE OF A (PERFECT) PLANET OBSERVER

AFTERNOON

- TCS/TELESCOPE

How to start the TCS and telescope is very well described in the DFOSC manual ( chap 1 for normal operation,

chap 5 for problems).

Remember, in case of problem, that telescope must be at zenith to be initialised. If it does not move back to zenith

you must do it by hand (and carefuly) as in DFOSC manual 5.3.3

The most frequent problems with "hardware" up to now have been:

=- Dome stayed in "Automatic" at the end of init: You must go to "manual", then "automatic".

=- The telescope does not go back to zenith from Dome flats position (it says that "previous preset not finished").

raise the "virtual handset" and click "guide".

=- The dome does not move (even when in "automatic" mode and initialised): find the miracle switch p49, 5.3.7

=- Adapter does not initialize. try as many times a needed as in 5.3.12

-BIAS COMPUTER

read the manual . . . We use bias in mode "run <file>.bach" with a set of xxx.bach files in /data/dfosc or in command modes, depending on the taste of the observer.

WorkSpace 1 is for BIAS itself, in WS#2 we open 1 emacs per useful .bach file.

- W1D5PL (PC)

=- Clean $RAW_IMAGES from last night images. They should be moved to a directory

with a code date such as may10, june23. There should not be any ".mt" file left.

( if so, they are images not understood by the system: worth a glimpse with dcsh_lsInfo )

Cold Start:

> login as guest, then type tcsh then source .planet

HOME is /home/guest/PIPELINE/CCSH

> If needed dcsh_OpenWindows opens 6 windows:

- 2 windows "mistyrose" traditional for flats and import. One in $RAW_IMAGES, the other in $IN_IMA

- 2 windows "papayawhip" for pipeline itself.

- 1 small window "lemonchiffon" to spy the import directory:

> su - guest (+Love4*S)

cd /scratch/guest/DFOSC_FASU/<date> <date> Ex. 2004-06-12

it is where images from dfosc arrive. "ls" will show you images pending for import.

- 1 small window "lemonchiffon" too, where I propose that you type:

cd $SCRIPT_CODE ; tail -f MainSummary.txt

offset and cross write one line/event in this file, as processed or crashed.

This gives you a fast way to keep an eye on the processing.

BIAS & FLATS

TAKING

> switch on the lamp (it must warm up a bit). level is 40.

> Telescope to FFscreen, open both miror cover and sky baffle

> make a test image with time 10s for R dome flats. level should be ~32000

(if not, check lamp in the dome, or miror cover...). One bulb only is on.

there are spares in the closet in the contol room)

> launch domeFlatR.bach We want 20 flats minimum

check in the "guest" window the appearance of the directory of the day, and check for

"lost" dfsc images.

Lost images: You can ftp them from INS1D5 to WRL1D5:

- Open a terminal window in INS1D5, or better make an:

> xterm -bg maroon -fg lemonchiffon -fn fixed -sl 5000 -e tcsh -l

> cd /data/dfoscobs

> ftp w1d5drl login: guest

ftp> cd /scratch/guest/DFOSC_FASU/<date>

ftp> put dfscxxxx.mt dfscxxxx.fits

> idem for V flats (time ~65 s)

After Dome flats are done, go to zenith, close mirror cover and sky baffle, and take 20 bias minimum

normal bias level is ~280

When finished, open the dome, side ports, mirror cover and sky baffle, side ports, and switch on the fans

on the rack in the control room , weather permitting.

Remember: THE MAIN MIRROR COVER MUST BE CLOSED when opening/closing the Dome.

FLATS & BIAS PREPARATION

Follow the "README_BIASFLATS", i.e.

- In the Import pink window, import images with

> dcsh_Import

> csh_prep_lists

> csh_bias_flat bias22jun flat22jun

Check the newly prepared flats. this is mandatory: it's a mess if you start processing with bad flats !

> cd $BIASFLAT

> ds9 bias22jun.fits

> ds9 domeFlat22junRf Vf If .... (must be ~1.0 within some %)

You are done. Have a good dinner, come back in time !

BEGINNING OF NIGHT.

One may start as soon as sol @ -12 deg.

MeteoMonitor. We must use mozilla on w1d5drl. As the computer is used for pipeline, etc,

it might be convinient to have meteo on the TX terminal 123.25 near the printer.

Open the TX session on ins1d5, then on w1d5drl, open a window on the TX with:

guest@w1d5drl> xterm -display 134.171.123.25:0.0 &

In the newly opened window on the TX, run mozilla. www.ls.eso.org will give you the link to MeteoMonitor.

On the telescope: Make a "CorCo"

CorCo is mandatory: our stamp on the CCD is not at the center !

Read once the corco text in the dfosc manual. The aim of the game is to "bias" the telescope

pointing, so that, for given RA,DEC, the star shows in the middle of our stamp.

You can use test images on the CCD, Guideprobes in park position, or "center" the guide

probes (they blind the telescope, but see what it would see), and use the TV camera.

Using TV is faster. I have written on the TV screen a cross where must stay the star when it is

well positionned on the planet stamp.

Choose a bright star around zenith. point the telescope. Don't forget equinox 2001.5

Set the guide probes to center field, wait it is done. Be careful when rising the voltage:

The star must appear as a HUDGE thing showing at very low voltage. You may burn the system

at that stage. The game is to put it on the cross on the TV monitor.

Generaly, the star does not show on the TV at first. I use to offset the telescope:

press "offset" on the hanset, then:

~ 12 times RA- DEC-

~ 5 times RA- (for the standard values of offset : 10 arcsec in the Tellescope Constants.

The star or it's light should then be visible in the left/down corner of the TV.

go on with the handset until on the cross

WARNING: the telescope is a heavy thing. When you press an offset, it displays

"offsetpr", and when done: "offset" again. You must wait untill it has finished before offsetting

again. Do not try to go too fast !

We use to refine the position with a few test images ( 0.1 s test): the target will lay within

~ 50 pixels from the center of the CCD (depending on elevation), and this is very

useful to track the lenses on first observations of alerts.

If the star does not show on the TV, you may try to see it on the ccd:

- park the guide probes

- make a test image of 0.1 s the star should be guessed in the upper right corner.

use RA- DEC- to drive it to the center of the ccd. This method is in fact more precise than the

cross on the TV monitor.

- If you are lost, even after several "preset" and tries, you can "cold start" again by going to

zenith, then initializing the telescope only to forget previous CorCo, and trying on a new star..

OBSERVING

It is: acquiring images at the telescope, importing and flatfielding them, and running the pipeline on them.

DURING THE NIGHT

AT THE TELESCOPE

Of course, your priority list is ready. Use catal/planet.cat for the stars. A good way not to waste time looking for a guide

star is to record the values for guide probes. One can prepare the object as well as G.P. values, and set all as soon as

the previous image is exposed.

The GP values for a number of alerts are in /catal/planet_guide . You can "cat" this file in a narrow window on the side

and cut/paste the values to the GP input.

Warning: the TCS checks from time to time if the catal directory has been modified. If it has, it will desable the

"cat Select" button : you cannot any more select a catalog, unless rebooting the TCS (an re-initializing all, and CorCo).

* Do not forget to check the serrurier temperature and to update the focus * (BIAS>> telfocus XXXX )

PROCESSING:

Processing is building the light curves, in archive files ( in $ARCHIVE/2004/), but also checking the quality of data, and

detecting and advertizing for any strange behaviour - caustics etc..

I use to loop on the following commands:

IN AN "Import" (pink) WINDOW:

> dcsh_loop_objects bias 22jun Domeflat22jun

This programs loops on Import and flatfielding. It is basicaly made of:

> dcsh_Import

> csh_flatfield bias22jun Domeflat22jun (this copies the files with planet names in $IN_IMAGES)

> sleep 120 sec

IN ONE WORKING WINDOW

> ccsh_loop_crs (or loop_off for offset )

We loop on processing of .fits images written in $IN_IMAGES by dcsh_loop_objects.

To stop the loop at end of current image, use (in any other window):

> ccsh_stop_loop

IN THE OTHER WORKING WINDOW

Checking archives, tuning refs stars, and any other work. See READ_ME_CCSH .

> ccsh_plotref <ArchiveName> Ex: ZOB04250R will show the lightcurve, and the refs stars.

> ccsh_anarch <ArchiveName> will show the same thing, but with numbers and a number of checks.

A point with Lens Mag 0 means that the lens has not show up: it should be removed

and the archive re-created.

Use ccsh_setbad <image> Ex: ZOB04250R067 to yank the point.

Use ccsh_arch <ArchveName> 1 N C to recreate archive. N : number last image.

> ccsh_refs <ArchiveName> can propose you some refs if the current ones happen to have big

dMag or sMag.

To change reference stars, template (reference frame) number, lens number, you must edit the data file

that all ccsh routine use to get information. There is one per alert, in $ARCHIVE/2004/ with name

arch<AlertName>.dat Ex: archZOB04250.dat (note: no "colour" letter: all colours in the same file).

Edit the archXXX.dat file, then enter the new values in the archive with:

> ccsh_writepl <ArchiveName>

And check agian (and again).

FROM TIME TO TIME, IN ANY (other) WINDOW, send results to HomeBase

> dcsh_Synchronize

Synchronize archives (i.e. sends all archives to planet@chianti.iap)

Update plots

remlog images

(the update of plots does not show immediately on the planet pages - needs a big couple of

minutes)

REPROCESSING

If you need to reprocess an alert ( changing template f.e. ), stop the processing loop; otherwise nasty

interferences are probable. Reprocessing is difficult, because the PC has limited power, and the night may

end very late.

to reprocess a field ( i.e. ZKB04033R ) in a row after template has been run, one may

- set the "LensedStar" to 0 in the arch.dat file

- ccsh_unwait/crash the wanted images

- launch ccsh_fld_crs ZKB04033R

which will process only this alert.

- As soon as the new archive has been created, restart the loops.

At END OF NIGHT:

CLOSE THE TELESCOPE

STOP LOOPS once the last event is processed

> dcsh_SyncArchives

COUNTS for Planet Report

After the last image has been imported (no need it is processed )

use dcsh_counts. This function works in the following way:

> dcsh_counts night26jun.cnts nights25jun.cnts

will: - create a file night26jun.cnts with all the counts now, i.e. at the end of night 26 jun.

- compare and substract the counts of night25jun, kept in nights25jun.cnts (or earlier if bad weather)

- print a list of what to put in the planet night report

Warning: night26jun counts are calculated if the file does not exist. If the first arg is an existing file, dcsh_counts

will use it . . . In case of error, better delete it and recreate it.

This system has to be initialised ... You can create a counts file any time with your actual counts at that time

( preferably at end of night ) with:

> ccsh_WSInfo 0 . | tee nightxxyy.cnts

USEFUL in case of problem:

> dcsh_lsInfo <dir> or image*

works as an "ls" but gives some extra info's on the images (dfsc, OBJECT, UT, MJD, PL_DATE)

GENERAL:

> BIAS data space must be cleaned for dfscxxx.mt only when advised to do so (for Eso archiving )

> Same: do not delete dfscxxx.mt files on INS1D5 (but you may erase the .bdf files created by Midas).

> An archive of all taken images is kept w1d5drl in : /data/DFOSC/AutoCd/BigDir/

> check the space left on drl disk /scratch (df -k). Actually, there is much less space than expected

and you may run out of space rapidly. I had no time to find why.

SyncArchives:

If dcsh_SyncArchives does not work after several tries, it is generaly because chianti.iap is not listening.

One may try to ftp by hand the archives ( or at least the urgent ones ):

> cd $ARCHIVES

> sftp planet@chianti.iap.fr

> password is 1venus@

> cd /home/chianti2/planet/planet/NotPublic/Archives/2004

> put ...... or mput .....

NEW ALERT - BUILDING AN ARCHIVE

-------------------------------------------------------------------

For clarity, we are going to build the archive in Red for alert OB-04-382 : ZOB04382I

We have a few images already, which have been put in the "wait" state by the Import system (as there was

no Template (reference frame) set for this object.

> ls $IN_IMAGE/ZOB04382I*

listed: ZOB04382I001.fits.wait ZOB04382I002.fits.wait ZOB04382I003.fits.wait ZOB04382I004.fits.wait ZOB04382I005.fits.wait

in $IN_IMAGE, images can be in different states:

.fits.wait No template available.. waiting

.fits.processed successful processing

.fits.crashed offset or cross ended badly

.fits.bad set by hand: bad point to be ignored

1) stop the pipeline (ccsh_loop_cross) to avoid nasty interferences, using

> ccsh_loop_stop

2) decide of the template frame.

> ccsh_fwhm ZOB04382I

will list all the images, with fwhm in pixels. Choose an image with good seeing. Here we choose image 004

3) Build a new $ARCHIVE/arch.ZOB04382.dat file by copying the TEMPLATEarch.dat and editing.

- set ReferenceFrame to 4 NoArchiving

NoArchiving tells some ccsh not to try archiving: you want to create the archive yourself.

- set NumLensStar to any non zero value. traditional is 666 we will set it at the end.

- Save it

4) build template

> ccsh_unwait ZOB04382I004

ccsh functions run only on images with end ".fits". Ours are ".fits.wait" remove wait state.

> ccsh_template ZOB04382I004

check there are a majority of type 1 (good) stars.

> ccsh_cross ZOB04382I004

process the template with itself as reference.

Now the template image is labelled: ZOB04382I004.fits.processed

5) Process the other images

> ccsh_unwait ZOB04382I will unwait all the 382I images in one turn

> ccsh_fld_crs ZOB04382I This function is a modified "loop": it will process all images .fits

for the given "field". "crs" stands for "cross". check Quality.

Now you should have all images either .fits.processed or .fits.crashed.

If some are crashed, you should try another template frame... or look at the images and setbad.

6) Create archive with first values

> ccsh_arch ZOB04382I 1 6 C ccsh_arch <archName> firstImage LastImage C for create

At the end of the creation, ccsh_arch lists the reference stars that you set at random in the arcxx.dat

file. If you find the words: "Rejected stars x points", better to tune the reference stars rapidly.

7) Set the lensNumber.

Find the lens on the template image with

> ds9 $IN_IMAGE/ZOB04382I004.fits.processed

We use to write the coord in the ReferenceFrame field of the arch file, after a # for that to be ignored

REMOVE THE "NoArchiving" field too. a typical arch line would be:

R ReferenceFrame 004 # NoArchiving 45 @ 229 246

Search for the number of that star in the archive

> ccsh_searchXY ZOB04382I 229 246 3 # i.e. in a box of 3 pixels

Search will list one star in a sea of "*****". If not, increase the radius. If many... choose.

*********************************

******NumStar=6056 XRef=666 USE NumStar

******XRef=665 664 663 662

******NumDop=6800

X=228.860,Y=257.730

FluxRef=-9.49, FluxMean=18.39

SigRef =0.104, SigMean =0.049

TypRef=11 TypMean=711

*********************************

- Update the archZOB04382.dat file with the lens number

we use to put it as a comment in the RefFrame line too, for memory, so your arch.dat file looks:

I ReferenceFrame 004 # NoArchiving 6056 @ 229 246

I NumLensStar 6056

- Update the archive with the new values in arch file:

> ccsh_writepl ZOB04382I

8) Check the result !

> ccsh_plotref ZOB04382I

> ccsh_anarch ZOB04382I

You can now restart the ccsh_loop stuff i.e. restart processing the images

9) Tuning

ccsh_refs ZOB04382I

to get a list of possibly better refs. update arch file, run writepl, check the result.

HOW TO XMIT GRB IMAGES to ASTRO.KU.DK

-------------------------------------------------------------------

WHAT?

I put all the stuff in 1 directory/grb, under PIPELINE/GRBS . Of course the amount depends of what grb people want.

For GRB0624 (24june) : I have put there the dfsc images (raw), and the flatfielded results ZGRB04624 (the 04 is

needed in the object name to be recognized as an "object" for planet - Do not put it if you take full CCD images !

I propose to put there flats, bias, flats images, etc if wanted by grb people.

It is possible after to send the directory in one time.

SENDING GRB DATA

A) Non deflatted images. sending one by one.

I used to send them from guest <day> directory, as dfsc numbers are known.

> su as guest +Love4*S and go to /scratch/guest/DFOSC_FASU/2004-0x-xx

> scp dfscxxxx.fits.gz.done too@astro.ku.dk:/p4/too/upload/dfscxxxx.fits.gz

you are prompted for the passwd of "too" : 7hizgrbs (each time!)

One may use regular expressions.

B) Deflatted images. sending a batch of files.

One can put all images in one directory and t\send them with one command.

I store grb images in PIPELINE/GRBS, one dir per grb. Exemple for GRB040624 (24jun)

all moved to PIPELINE/GRBS/GRB04624

> (under planet account, this time)

> cd GRBS

> scp -r GRB04624 too@astro.ku.dk:/p4/too/upload/